Eh, it's just a fact that most startups fail. They tried, they ran out of money. Launching an IoT startup in 2012 in Australia was pretty ballsy. Total outside investment was only $1.7m over 3 years. Had they been overfunded like a stereotyped smoke-bloated Valley startup, maybe they would have had more runway.
And you think it's perfectly ok for a company to say, "hey sorry guys we wont be delivering what we promised you because we made bad decisions and spent all your money." Sure companies fail but this is ridiculous. A typical example of a heavily promoted company riding a wave of hype in to the gutter. I'm sure it's hard for the guys who built this company but they should be ashamed of not delivering what they promised 300 of their customers. I found their post rather cavalier and deserving of some kicking.
I'm not saying HN should be a place of ruthless criticism and shnarky comments but I'm also not in the camp that thinks it should all be cuddling and cooing either.
HN's values aren't "cooey and cuddly", they're "civil and substantive". That leaves plenty of room for criticism, so I'm not really seeing the worry here.
Although this is a shame it is yet another sign (for me) that the 'Internet of Things' is going to be DoA if we have to also rely on them for a backend cloud service. Most of these devices could be run in a completely decentralised way, such that if the company goes under, your device is hardly affected. e.g. I don't expect my light switches to stop working if the parent company gets acquired or shuts down.
There are real challenges to making distributed systems easier to build but I'm working with others on that c.f http://nymote.org and http://openmirage.org.
Completely agree. But of course using a cloud service means no need to run a server on your own network/device or have to punch holes in your router to have the device accessible whilst you are not at home. Requiring either of those wouldn't help IoT adoption outside of enthusiasts.
If you could give everyone their own 'personal cloud', without having them become sysadmins, then it's possible (of course, that's not easy but it's also why it's a valuable goal). The work under Nymote has a longer term vision so not ready to help people just yet -- some more detail at http://nymote.org/blog/2015/brewing-miso-to-serve-nymote/
In this instance, I bet sandstorm.io could come up with a transition for Ninja Blocks users. I hope they think about it.
I don't expect my light switches to stop working if the parent company gets acquired or shuts down
That's the simple case, where you are only using devices from a single vendor. If you - god forbid - own devices from two different companies and try to connect them in any meaningful way, your setup might suddenly stop working as soon as one of the following happens:
- Company A shuts down.
- Company B shuts down.
- The hub service that you used to connect Product A and Product B shuts down.
- Any one of the involved companies have a dispute with any other of the involved companies and cease cooperation/demand that their products not be used together until contracts are re-negotiated.
- Any of the companies makes a breaking API change and the other's don't have the time/money/motivation to update their software accordingly.
Lights can still be switched off/on without the ninja blocks cloud, it still works through your LAN even if internet is down. You lose the rules engine though. I like this feature on my NB. I use it for lamps around house, triggered from buttons on wall and in phone app. Bypasses cloud and the switch response time is very fast because of this. Hasn't missed a beat for a long time, the Ninja Block is a great little device.
For any of this IOT stuff, a reliable cloud service is essential. Apparently there are ways to have my own "mini cloud" for my Ninja Block, so I'll be following that up before the official one shuts down.
This is only an issue if you buy proprietary devices for the platform. As long as the individual components (lights, sensors etc) use standard formats (z-wave etc) you should be able to run them on any IoS platform.
This sucks. I wanted to buy one because it seemed to do what I want, but I guess they have decided that its better to go under than sell me one.
All I was waiting for was better support for existing services and devices.
Anyone behind YC want to go buy them to keep them alive? Its a massive waste to let the company go under when they have customers wanting to buy their product.
One thing I feel the NB never quite achieved was "state". It sort of worked, but sometimes not the way expected. It turns out that "state" - like a door open or closed, is not easy to keep track of reliably. Caution is needed for anything else in the chain eg "if door open for longer than 30 sec, sound alarm".
In the end, nothing beats a physical switch on the door (wireless battery operated), with hardware on/off states. Then the door can't be mistaken ever as open or closed due to mis-matches in sending the state, or whatever other interference cause the "door open" state to get out of sync. If Ninja Blocks had made their own special little purpose-built wireless sensors and switches rather than use aftermarket, then things like state could have been sorted out early on IMHO.
The supplied door switch for example that came with the NB, triggered the same thing whenever the pin was moved in or out. The switch had no concept of open or closed or on/off, it just blurted the same signal each time it was triggered. Attempting to build "state" from such trigger data is never going to be reliable.
Knowing without doubt that your garage door is open or closed when looking at the app on your phone, is not something you want your security device to be wishy-washy about. You want the truth. Ninja Blocks (in combo with the basic remote sensors) simply couldn't give you the truth, it could only give you a "pretty sure, quite likely, almost certainly the door is closed". This isn't good enough for security.
Ninja Blocks just became the poster child for cloud based hardware failures.
The product never "needed" to be cloud-based. The dashboard could have lived on the Beaglebone Black inside the main unit and from there it could have allowed us to hook it into IFTTT, Pushbullet, etc. without the need for Ninja Blocks to still be a part of it.
These sort of failures make the larger corporate offerings look even more sweet, and that is a terrifyingly dangerous prospect. The company might not go away, but the data they will collect on you will probably blow your hair back.
20 comments
[ 4.3 ms ] story [ 55.7 ms ] threadPlease don't post mean comments to Hacker News.
Your post reads like you are 9 years old. Welcome to the real world.
I'm not saying HN should be a place of ruthless criticism and shnarky comments but I'm also not in the camp that thinks it should all be cuddling and cooing either.
No; more than one thing is not ok.
HN's values aren't "cooey and cuddly", they're "civil and substantive". That leaves plenty of room for criticism, so I'm not really seeing the worry here.
There are real challenges to making distributed systems easier to build but I'm working with others on that c.f http://nymote.org and http://openmirage.org.
In this instance, I bet sandstorm.io could come up with a transition for Ninja Blocks users. I hope they think about it.
That's the simple case, where you are only using devices from a single vendor. If you - god forbid - own devices from two different companies and try to connect them in any meaningful way, your setup might suddenly stop working as soon as one of the following happens:
- Company A shuts down.
- Company B shuts down.
- The hub service that you used to connect Product A and Product B shuts down.
- Any one of the involved companies have a dispute with any other of the involved companies and cease cooperation/demand that their products not be used together until contracts are re-negotiated.
- Any of the companies makes a breaking API change and the other's don't have the time/money/motivation to update their software accordingly.
For any of this IOT stuff, a reliable cloud service is essential. Apparently there are ways to have my own "mini cloud" for my Ninja Block, so I'll be following that up before the official one shuts down.
All I was waiting for was better support for existing services and devices.
Anyone behind YC want to go buy them to keep them alive? Its a massive waste to let the company go under when they have customers wanting to buy their product.
In the end, nothing beats a physical switch on the door (wireless battery operated), with hardware on/off states. Then the door can't be mistaken ever as open or closed due to mis-matches in sending the state, or whatever other interference cause the "door open" state to get out of sync. If Ninja Blocks had made their own special little purpose-built wireless sensors and switches rather than use aftermarket, then things like state could have been sorted out early on IMHO.
The supplied door switch for example that came with the NB, triggered the same thing whenever the pin was moved in or out. The switch had no concept of open or closed or on/off, it just blurted the same signal each time it was triggered. Attempting to build "state" from such trigger data is never going to be reliable.
Knowing without doubt that your garage door is open or closed when looking at the app on your phone, is not something you want your security device to be wishy-washy about. You want the truth. Ninja Blocks (in combo with the basic remote sensors) simply couldn't give you the truth, it could only give you a "pretty sure, quite likely, almost certainly the door is closed". This isn't good enough for security.
The product never "needed" to be cloud-based. The dashboard could have lived on the Beaglebone Black inside the main unit and from there it could have allowed us to hook it into IFTTT, Pushbullet, etc. without the need for Ninja Blocks to still be a part of it.
These sort of failures make the larger corporate offerings look even more sweet, and that is a terrifyingly dangerous prospect. The company might not go away, but the data they will collect on you will probably blow your hair back.
IoT needs an open source solution, and fast.